Current Concepts Review

Selective dorsal rhizotomy in ambulant children with

K. K. Wang1,2 Introduction 1,3 M. E. Munger Primary neurological factors (e.g. , weakness, 1,4 B. P.-J. Chen balance, poor selective motor control) and secondary T. F. Novacheck 1,4 deformities (e.g. joint , bony deformities, joint dysplasia) both compromise gait and motor function in children with cerebral palsy (CP). Correction of second- Abstract ary deformities can be done via orthopaedic interven- Background Selective dorsal rhizotomy (SDR) is a surgical tion, often in the form of single event multi-level surgery procedure for treating spasticity in ambulant children with (SEMLS).1 However, this does not address the primary cerebral palsy (CP). However, controversies remain regarding neurological factors. A growing body of literature sug- indications, techniques and outcomes. gests that selective dorsal rhizotomy (SDR) is effective in Current evidence summary Because SDR is an irreversible improving one neurological factor – spasticity, which may procedure, careful patient selection, a multi-disciplinary ap- lead to improvements in function and gait. proach in assessment and management and division of the Spasticity is defined by increased resistance with increas- appropriate proportion of dorsal rootlets are felt to be par- ing velocity of movement and/or the presence of a spas- 2 amount for maximizing safety. Reliable evidence exists that tic catch. A number of interventions are available for the SDR consistently reduces spasticity, in a predictable manner management of spasticity. Broadly, these can be described and to a substantial degree. However, functional improve- as surgical versus non-surgical, temporary versus perma- 3 ments are small in the short-term with long-term benefits nent and focal versus generalized (Fig. 1). SDR is a form of difficult to assess. surgical, permanent and relatively generalized treatment option for lower extremity spasticity that reduces tone by Future outlook There is a need for high-quality studies uti- selective sectioning of lumbosacral afferent nerve rootlets. lizing long-term functional outcomes and well-matched However, much controversy remains regarding the use control groups. Collaborative, multicentre efforts are of SDR, how patients should be selected for the procedure required to further define the role of SDR as part of the and what outcomes can be achieved.4-6 Quality long-term management paradigm in maximizing physical function in data are lacking, and much of what is done in clinical prac- spastic CP. tice is not supported by strong evidence. In addition, nat- 7 Cite this article: Wang KK, Munger ME, Chen BP-J, ­Novacheck TF. ural history studies exist that show reduction in spasticity 8 Selective dorsal rhizotomy in ambulant children with ­cerebral and improvement in gross motor function can occur in palsy. J Child Orthop­ 2018;12:413-427. DOI: 10.1302/1863- the absence of intervention, making results of SDR diffi- 2548.12.180123 cult to interpret without well-matched control groups. This article aims to serve two purposes: 1) to explore the Keywords: cerebral palsy; selective dorsal rhizotomy; boundaries of our knowledge by providing a review of the spasticity; interdisciplinary care current literature; and 2) to share three decades of expe- rience from our institution on the use of SDR in helping children with CP. 1Center for Gait and Motion Analysis, Gillette Children’s Specialty Healthcare, Twin Cities, Minnesota, USA The focus of this review will be on the role of SDR in 2Department of , Royal Children’s Hospital, maximizing mobility and independence for ambulant Melbourne, Victoria, Australia children with CP, functioning at Gross Motor Function 3 Department of Orthopedic Surgery and Department of Health Classification System (GMFCS)9 levels I to III. A nonselec- Policy and Management, University of Minnesota, Twin Cities, tive dorsal-ventral rhizotomy may be performed in indi- Minnesota, USA 4Department of Orthopedic Surgery, University of Minnesota, viduals with more severe impairments, functioning at Twin Cities, Minnesota, USA GMFCS levels IV or V, in order to meet palliative care goals such as pain relief and improvement of caregiving.10 Since Correspondence should be sent to T. F. Novacheck, Center for this is a different procedure with different indications and Gait and Motion Analysis, Gillette Children’s Specialty Healthcare, 200 University Ave E, St Paul, MN 55101, United States. ­expectations, it is outside the scope of this review and will E-mail: [email protected] not be addressed here. CURRENT CONCEPTS REVIEW: SELECTIVE DORSAL RHIZOTOMY

Fig. 1 Tone management options in cerebral palsy. Tone management is only one aspect of the musculoskeletal care needs of children with : lever arm dysfunction and joint deformity are other important aspects and are not represented in this diagram. Within the tone management paradigm, options include non-invasive (green), injections (yellow) and surgical (red). Selective dorsal rhizotomy (SDR) is a form of surgical, irreversible and largely generalized option for tone reduction. Orthopaedic procedures for tone reduction include muscle/tendon lengthening and transfers (BoNT, ; ITB, intrathecal ; DBS, deep brain stimulation; SDR, selective dorsal rhizotomy).

History and rationale In recent years, some surgeons have adopted a more lim- ited at the level of the conus as advocated In preterm infants, the deep periventricular white matter by Park and Johnston.16 Although technically more chal- is metabolically susceptible to injury. Damage to this area lenging for identification of rootlet levels, proponents of typically leads to reduced inhibitory input into the spinal the conus approach argue that it results in a smaller scar, interneuron pool from descending neural pathways. This reduced iatrogenic spinal instability, decreased postoper- results in excessive alpha motor neuron activ- ative pain and quicker recovery.17 ity and spasticity. Afferent dorsal root input into the spinal Following his relocation to the United States from South interneuron pool has a further net excitatory effect on the Africa, Peacock helped popularize SDR in North America for on the alpha motor neurons. Dorsal rhizotomy reduces the treatment of children with spastic CP.18,19 Understand- the amount of excitation of alpha motor neurons and ing that children with spastic CP are heterogenous, Peacock thereby reduces spasticity (Fig. 2).11 helped establish a set of selection criteria for SDR to maxi- Although the first dorsal rhizotomy was performed mize safety and efficacy. These criteria were adopted by the more than 100 year ago,12 the procedure’s evolution multidisciplinary team of orthopaedic surgeons, physiatrists since then has taken several major turns. Foerster’s initial and neurosurgeons at our institution in the 1980s. In the method of non-selective rhizotomy fell into disuse for more three decades since, our protocol for the care of children in than five decades because of the adverse consequences of the SDR pathway has stayed relatively constant. The original excessive de-afferention. In the 1960s, the procedure was criteria were incorporated with pre- and postoperative 3D revised by sectioning only a fraction of the rootlets based gait analysis (3DGA), a multi-disciplinary team approach, as on the patient’s preoperative function,13 and then further well as intensive postoperative inpatient rehabilitation. refined intoselective dorsal rhizotomy by electrically stim- ulating the rootlets intraoperatively and measuring the 14 electromyographic response before sectioning. In South Patient selection for SDR Africa, Peacock shifted the site of rhizotomy from the conus medullaris region to the cauda equina.15 This makes root- Although similarities exist between patient selection cri- let identification easier in an effort to avoid bladder dys- teria of various institutions, there is no consensus. Several function, one of the complications noted by Fasano et al.14 factors contribute to the lack of uniformity. Firstly, there

414 J Child Orthop 2018;12:413-427 CURRENT CONCEPTS REVIEW: SELECTIVE DORSAL RHIZOTOMY

Fig. 2 Spasticity reflex arc schematic diagram. Muscle stretch stimulates dorsal (afferent) sensory nerve rootlets, which in turn has a net excitatory effect on alpha motor neurons within the spinal cord. The spasticity arc is completed via the hyperstimulated efferent ventral motor rootlets. During SDR, individual dorsal rootlets are isolated and stimulated with a Peacock probe. Those that produce inappropriate activation are sectioned. Illustration used with permission, courtesy of Timothy T. Trost.

Table 1 Selection criteria for selective dorsal rhizotomy (SDR). These characteristics constitute the ‘ideal’ candidate for SDR. The perfect patient that fits all preoperative categories is rare and so discretion is necessary based on an individual’s overall picture

Features ‘ideal’ for SDR Notes

History Age: 4 to 10 years See discussion Birth events: hypoxic ischemic encephalopathy More likely associated with periventricular leukomalacia Birth events: prematurity < 36 weeks gestation Spastic diplegia More likely to benefit than quadriplegia (Kim 2006)66 Absence of other causes of spasticity E.g. hereditary spastic paraplegia

Examination Functional level: GMFCS I to III Ambulant cerebral palsy (Josenby et al 2012,38 Dudley et al 201337) Spasticity: Ashworth Scale 2 to 4 Multi-level spasticity No mixed tone - Hypertonia Assessment Tool Identify underlying Strength: at least anti-gravity hip flexors Mean medical research council (MRC) grade > 3 for lower limb muscles Selective motor control Complete or partial ability to isolate movement at hips, knees and ankles Absence of contractures Contractures will not resolve post SDR Cognition Able to participate in rehabilitation

Gait analysis Optimum patterns on kinematics See discussion Dynamic Motor Control Index Calculated from data Energy efficiency > 200% of normal controls

Imaging Periventricular leukomalacia ‘Classic’ hypoxic ischemic encephalopathy of prematurity No involvement of /thalami Associated with dystonia are a large number of preoperative parameters that are selection criteria. However, much like other paediatric considered, as highlighted in a recent systematic review interventions that aim to change a condition’s natural his- on this topic.20 The ‘ideal’ child that perfectly fits all pre- tory far into the future, long-term outcomes that extend operative categories is rare and so discretion is often nec- into adulthood are difficult to obtain. Thirdly, most clini- essary based on an individual’s overall picture. Secondly, cians agree that a major goal of SDR in ambulant CP is feedback from appropriately designed long-term out- to improve quality of gait and mobility. Therefore, 3DGA come data is crucial for defining the optimal preoperative should form a critical component of preoperative selection

J Child Orthop 2018;12:413-427 415 CURRENT CONCEPTS REVIEW: SELECTIVE DORSAL RHIZOTOMY

and postoperative evaluation. However, 3DGA is not used will have scores between 2 and 4 at multiple muscle lev- by every centre that performs SDR and is rarely included els in the lower extremities. Other options for quantifying in studies on indications or outcomes. spasticity include the Tardieu scale29 and instrumented The criteria for the ‘ideal’ SDR candidate at our institu- measures.30 Dystonia is defined as sustained or intermittent tion are based on Peacock’s pioneering work15 (Table 1). involuntary muscle contractions that cause twisting and Short- (18 months) and long-term (13 years) results at our repetitive movements or abnormal postures.31 The HAT is a institution based on these criteria have been published qualitative tool shown to have moderate to substantial reli- previously.21,22 Our criteria here share similar items to those ability in differentiating between spasticity and dystonia,26 published by other institutions, such as Oswestry23 and although validity is not confirmed.32 Recent efforts to use Great Ormond Street24 (Appendix A). Exceptions include dynamic electromyography (EMG) on physical examination our use of 3DGA and standardized assessment tools of and during gait have also been attempted, but differentiat- spasticity, such as the Ashworth scale25 and the Hyperto- ing between the two remains challenging.33 nia Assessment Tool (HAT).26 Several items on the criteria list in Table 1 warrant fur- Selective motor control ther discussion. Individuals with CP have varying degrees of impairment in selective motor control. This can manifest as primitive Age movement patterns (mass flexion/extension), and ‘pat- Spasticity is a major limiting factor of function in early and terned’ activation of muscle groups when attempting to middle childhood. In later childhood and adolescence, perform simple motor tasks. Compromise in motor con- surgical correction of lower extremity lever-arm deformi- trol is not known to improve following removal of spastic- ties forms an essential aspect of management.1 Typically, ity with SDR. Therefore, the ‘ideal’ candidate should have both hypertonia and secondary deformities need to be relatively good baseline selective motor control; 3DGA addressed for optimal results, although each may be more can help differentiate between gait abnormalities due to appropriate at different age ranges. spasticity or poor selective motor control.34 Since one of the concepts of SDR is to train a new gait pat- Motor control can be quantified statically or dynamically. tern in the absence of spasticity, SDR is believed to be more Physical examination can measure static selective motor effective if performed in early childhood. If performed too control using a scale of 0 to 2, indicating observation of late, secondary contractures may develop that could limit only patterned movement (0), partially isolated movement effectiveness. Classic studies reporting benefits of SDR were (1) or complete isolated movement (2) for each muscle.35 performed in young children generally between four and Motor control can be dynamically quantified from analysis ten years of age.27,28 MacWilliams et al6 showed that older of muscle synergy complexity using surface EMG during patients who underwent SDR between the ages of ten and 3DGA.36 As measured by the Walking Dynamic Motor Con- 20 years experienced functional declines compared with trol Index, better motor control is associated with better those who did not undergo SDR. Additionally, it is difficult functional gains following either SEMLS or SDR.36 to clearly identify candidacy for SDR in children who are younger than four years. Gross motor function is known to Baseline gross motor function be dynamic in the young child with CP, typically improving Authors have reported that children who are GMFCS I to up to the ages of six or seven years, with the steepest increase III are more likely to benefit than those who are GMFCS 8 occurring in the first three to four years. Cooperation with IV to V.37-39 This partly forms the rationale for includ- physical examination and 3DGA for the very young child is ing GMFCS I to III as a selection criteria at our institu- also difficult. These factors contribute to reduced reliabil- tion, as well as in most studies reported in literature.20 ity in measuring baseline preoperative functional level and However, some institutions rarely recommend SDR for therefore difficulty in projecting long-term outcomes. children who are GMFCS I,24 with the concern that the benefits gained may be too small to justify potential Spasticity versus dystonia risks. In GMFCS III children, concerns regarding poten- SDR only treats spasticity. Therefore ideally, spasticity should tial weakening effects of SDR can also limit selection. be the primary impairment in any candidate for SDR. In real- Physical examination and 3DGA are important for dis- ity, it is common for spasticity and dystonia to be co-present tinguishing between those children who rely on spastic- in varying degrees. Even though it is critical to differentiate ity to maintain antigravity strength and those who are between the two, there is no objective measure universally more impeded by their spasticity. Whilst we do believe agreed upon for either entity, especially dystonia. One mea- that GMFCS II children often form the most ‘ideal’ can- sure of spasticity is the , which is a didates, the risk/benefit profile of SDR is heavily influ- scale from 1 (normal) to 5 (rigid).25 An ‘ideal’ SDR ­candidate enced by the percentage of rootlets sectioned. Around

416 J Child Orthop 2018;12:413-427 CURRENT CONCEPTS REVIEW: SELECTIVE DORSAL RHIZOTOMY

25% to 40% of rootlets are routinely sectioned during SDR at our institution, which is low comparatively in literature. The higher the number of rootlets sectioned, the less ‘selective’ the rhizotomy procedure is, and the greater the risk of inducing weakness. Yet despite a rel- atively lower percentage of rootlets sectioned, we see no compromise in spasticity reduction achieved at our institution as assessed on modified Ashworth scale and long-term follow-up.21,22 Therefore, we believe that with careful selection of patients and judicial sectioning of rootlets, a proportion of children who are GMFCS level I or III can also benefit from SDR.

Kinematics on 3DGA The use of 3DGA kinematics can help objectively identify factors in a child that are either favourable or non-favour- able for SDR.34 As mentioned above, a pattern of ‘mass flexion/mass extension’ movement on the sagittal plane in the hip and knee suggests reduced selective motor con- trol (Fig. 3). Children with dystonia will also have kine- matic traces that vary significantly between individual gait cycles (Fig. 4). These factors are non-favourable for SDR, for reasons stated earlier. Since tone reduction is the goal of SDR, identification of kinematic patterns consistent with multi-level spasticity is critical. In the lower extremity, muscles most affected by spasticity cross more than one joint, and dynamic stiffness in the affected joints can be demonstrated objectively on 3DGA (Fig. 5). Findings include: – a ‘double bump’ pattern on sagittal pelvic kinemat- ics suggests the presence of underlying psoas or hamstring spasticity; – slowed, reduced and delayed knee flexion in early swing is consistent with rectus femoris spasticity; – reduced knee extension in late swing associated with posterior pelvic tilt suggests hamstring spasticity; – early ankle plantar-flexion on initial contact sug- gests spastic loading response in gastrocnemius muscles; – muscle-tendon modelling derived from joint kinematics can also show reduced length and velocity (Fig. 6). However, restricted range of movement can also be caused by established soft-tissue contractures which Fig. 3 Kinematic traces of ‘mass flexion (Flx)/extension (Ext)’. would not be expected to improve with SDR. Soft-tissue Mass flexion-extension is a primitive movement pattern suggesting reduced selective motor control. This can be seen lengthening procedures may be necessary to address typically between the sagittal hip, knee and ankle traces during these prior to SDR. Although it may be possible to tease swing. Flexion and extension are abnormally synchronized. out some of these findings in a detailed physical examina- Solid blue line represents onset of mass flexion (dorsiflexion in tion, spasticity is a problem of movement, and therefore ankle) and dotted blue line represents peak flexion in swing. Red traces: left limb. Green traces: right limb. Long green/red vertical a dynamic assessment of the child using 3DGA should lines: ipsilateral limb foot-off. Short green/red vertical lines: provide the best means of quantifying pathology and contralateral limb foot-off. assessing candidacy for SDR.

J Child Orthop 2018;12:413-427 417 CURRENT CONCEPTS REVIEW: SELECTIVE DORSAL RHIZOTOMY

Fig. 4 Kinematic traces in dystonia. Uncontrollable movements in dystonia results in large cycle to cycle variations between individual cycles. This individual also walks with plantarflexed gait, foot drop in swing and crouch (Pos, posterior; Ant, anterior; Ext, extension; Flx, flexion; Pla, plantarflexion; Dor, dorsiflexion).

Neuroimaging formation and ventriculomegaly with thin corpus callo- sum. In children with these findings, CP is more severe To identify the ‘ideal’ candidate, brain magnetic reso- and other factors become more important in limiting nance imaging (MRI) is essential for determining the mobility than spasticity alone, making outcomes of SDR presence of periventricular leukomalacia (PVL), and the less predictable. Injuries to the basal ganglia, absence of injury to basal ganglia, brainstem or cerebel- and are characteristic of hypoxic ischemic lum (Fig. 7).20 PVL compromises supratentorial influences injuries, more typical in full-term births.40 These MRI pat- to the spinal neuronal pool, resulting in loss of inhibition terns are associated with dystonia and are therefore not to the spinal reflex arcs. PVL is the most common MRI find- an ‘ideal’ feature in selection for SDR. Screening MRI of ing in CP and is seen in over 70% of children with spastic the spine are also routinely ordered prior to proceeding diplegia.40 The highest risk of PVL occurs after birth prior with SDR to rule out other anomalies of the neural axis. to 32 weeks’ gestation,41 and hence prematurity is also usually considered part of the selection criteria. Smaller lesions of PVL are limited to the trigonal regions of the Surgical technique ventricles and primarily affect the lower limbs. With increasing severity of injury, larger areas of periventric- Today, most centres perform SDR using variations of ular white matter are affected, leading to cavitation, cyst the technique described by Peacock.15,42 The patient is

418 J Child Orthop 2018;12:413-427 CURRENT CONCEPTS REVIEW: SELECTIVE DORSAL RHIZOTOMY

Fig. 5 Kinematic pattern of predominantly underlying spasticity affecting gait. ‘Double bump’ pelvis, slow and delayed knee flexion in early swing, reduced knee extension in late swing and early ankle plantarflexion in stance all reflect spasticity in underlying muscles. The involved muscles are often those that cross more than one joint (rectus femoris, hamstrings, gastrocnemius) (Pos, posterior; Ant, anterior; Flx, flexion; Ext, extension; Pla, plantarflexion; Dor, dorsiflexion). positioned prone and needle EMG sensors are positioned lower percentage of between 25% and 40% is routine, and to monitor representative muscles of each myotome, which near uniform normalization of spasticity is still achieved as may include psoas, vastus lateralis, tibialis anterior, pero- measured by our long-term outcomes via the modified neus longus and gastrocnemius. The surgical approach is Ashworth scale.21,22 In addition, a lower percentage of root- most commonly via a laminectomy or osteoplastic lamino- let resection may reduce risks of side effects of excessive tomy from L1 or L2 to S1, which allows excellent visual- de-afferention such as weakness and sensory abnormali- ization of the cauda equina and rootlets.11 The single-level ties. Bladder dysfunction is avoided by identifying sacral laminectomy at the level of the conus is a technically more rootlets responsible for bowel and bladder control using challenging alternative but has the advantages associated anal sphincter sensors and preserving them. 16 with a minimally-invasive approach. Following laminec- Irrespective of the surgical approach used or the exact tomy, the plane between ventral (motor) and dorsal (sen- percentage of rootlets divided, it is well established that sory) roots is identified, and ventral roots are protected SDR is effective in spasticity reduction.43 What remains throughout. controversial is whether this translates into long-term Dorsal roots are then separated into individual myotome functional advantages over other management strategies levels. Systematic stimulation of rootlets is carried out with in tone reduction. Current efforts are now focused on a Peacock probe at threshold amplitude, usually with a fre- answering these questions. quency of 50Hz. Rootlets with abnormal response are sec- tioned. Responses that are considered abnormal include those that are incremental, clonic, multiphasic, sustained Outcomes of SDR or spread to three or more adjacent levels or to the oppo- site leg.21 The total proportion of rootlets resected in some Interpreting literature on the outcomes of SDR must take institutions exceed 40%.11,24 However, at our institution, a into account several factors. First, the surgical intervention­

J Child Orthop 2018;12:413-427 419 CURRENT CONCEPTS REVIEW: SELECTIVE DORSAL RHIZOTOMY

carried out is a permanent one, and will affect the child changes in gross motor function and spasticity that may into their adolescence and adulthood. Therefore, long- occur with age. Well-matched control groups are there- term findings are particularly important. Second, the fore critical in comparing SDR with other management natural history of CP should be considered, including modalities. Third, in ambulant children with CP function- ing at GMFCS levels I to III, outcome assessment of SDR should include quantitative measurements of gait qual- ity in the form of 3DGA. However, routine 3DGA is not yet the standard of practice. Fourth, global assessment across multiple domains of the International Classifica- tion of Function (ICF) is sparsely reported. Even when ICF is reported, the effect on outcomes that family sup- port, income or other social factors may have are often ­unmeasured or ­underestimated.

Spasticity A relatively non-controversial outcome of SDR is its effect on spasticity.43 Several short-term randomized trials,27,28,44 as well as long-term cohort studies,22,39 have confirmed spasticity reduction following the procedure. However, this information should be understood in light of evidence shed on the natural history of spasticity in CP. Spasticity in the gastrocnemius muscle, as measured longitudinally by a dichotomized Ashworth scale, reached a peak at age four, followed by a gradual reduction up to the age of 12.7 Since most SDR procedures are performed between the ages of four and ten years, the implication is that the natural history of spasticity at least partially accounts for the improvements seen post-SDR. However, in our long- term study comparing children who underwent SDR with those who had alternate tone management strat- egies, SDR resulted in uniform reduction of tone on the modified Ashworth scale in all muscle groups down to a normalized score of one, whereas those in the compari- son group only showed a partial reduction in spasticity. Fig. 6 Hamstring length. Musculotendinous length modelling These findings strongly suggest that SDR produced more can be performed given known muscle insertions and joint positions. Spasticity is associated with short hamstring length additional reduction in spasticity than can be accounted and slow velocity, when matched for walking speed. These are for by natural history alone or by other tone management expected to improve following selective dorsal rhizotomy (SDR). strategies.22

Fig. 7 Brain MRI in periventricular leukomalacia (PVL). The ‘ideal’ candidate for selective dorsal rhizotomy will have isolated PVL (red arrows).

420 J Child Orthop 2018;12:413-427 CURRENT CONCEPTS REVIEW: SELECTIVE DORSAL RHIZOTOMY

Dystonia required, such as anti-spasticity injections and soft-tissue 22 One concern regarding SDR is that it may unmask dys- orthopaedic surgeries. tonia once spasticity is corrected. Isolated incidents of When stratified according to preoperative GMFCS this occurring have been reported for children with more level, authors have reported that children who are GMFCS severe impairment.10 This concern can be minimized by I to III are more likely to benefit than those who are 37-39 the careful selection of SDR candidates who show mini- GMFCS IV to V. This forms the rationale for including mal signs of dystonia or mixed tone. As suggested previ- GMFCS I to III as one of the selection criteria. It is possi- ously, the HAT tool can be used to identify dystonia, but ble that for a select few, SDR may allow a small improve- at present, there is no objective and quantitative measure- ment in GMFCS level itself. While GMFCS is not intended ment. Since there is no evidence to suggest SDR reduces as an outcome measure, some studies have reported 38 dystonia, alternative surgical strategies such as intrathecal changes post-SDR. Josenby et al found that seven out baclofen, ventral rhizotomy or deep brain stimulation may of 29 patients had an improvement of GMFCS level, and be required to address refractory and disabling dystonia. none had deteriorated, but no comparison group was available. While our long-term study had a comparison Functional outcomes group that was matched on a variety of baseline param- eters, few individuals in each group had both a baseline As mentioned previously, studies on long-term functional and follow-up GMFCS level measured. Five out of nine outcomes of SDR are limited by small sample size, hetero- children who underwent SDR had an improvement in geneous outcome measures, and lack of control groups. GMFCS level at long-term follow-up while only one dete- Available evidence suggests that SDR confers a modest riorated.22 The control group had neither improvement functional advantage. A meta-analysis of three short-term nor deterioration. (9 to 24 months) randomized control trials (RCTs) in 2002 found that SDR combined with improved Gait analysis Gross Motor Function Measure (GMFM) more than phys- ical therapy alone.43 However, when the included stud- To our knowledge, three studies to date have inves- ies are looked at separately, only two supported small tigated the effect of SDR on gait quality using pre- and functional improvements.27,28 The RCT by McLaughlin et ­postoperative quantitative gait analysis. Two of these stud- 48,49 al,44 which had the longer follow-up (24 months) did not ies used 2D gait analysis (2DGA), and only one used 22 49 show improvement in function, although a reduction in 3DGA. Subramanian et al found improvement in some spasticity was still seen. Observed short-term improve- 2DGA gait parameters at one and three years post-SDR, ments after intervention in a young child must also be but this improvement was not maintained at ten years. 48 placed in the context of gross motor function improve- In contrast, Langerak et al found that improvements in ments that are naturally expected up to the age of seven range of movement, cadence and step length persisted years.8 20 years post-SDR. At our institution, 3DGA is performed A more recent systematic analysis that included cohort routinely pre- and postoperatively for major interventions. studies with more than five years of follow-up found We reported our findings comparing matched groups that SDR conferred a modest improvement in ICF body of patients who underwent treatments with and with- 22 structure and body function domains, but had no influ- out SDR. At a median follow-up of 13 years, we found ence on ICF activity and participation domains.45 Of the that both SDR and control groups improved in terms of studies that included a comparison group, Daunter et al46 Gait Deviation Index (GDI), mean dynamic knee range of showed that SDR was associated with a lesser decline in movement and equinus in stance. However, the non-SDR gross motor function and reduced hours of daily assis- group had a greater improvement in GDI compared with tance required. However, changes in function were self- the SDR group at the time of long-term follow-up, likely rated and it is uncertain whether groups were matched at the cost of an increased number of orthopaedic inter- on baseline spasticity levels. Bolster et al47 showed that ventions. compared with reference centiles stratified by GMFCS Pain level and based on GMFM function, no children showed deterioration of function by more than 20 centiles, and a Although not a common manifestation of CP during child- portion showed significant improvement in GMFM scores. hood, the prevalence of pain in adults with CP is higher Munger et al22 found that although patients with SDR than in the general population.50,51 Some authors have were not significantly different to a well-matched con- hypothesized that persistent muscle spasticity could be trol group in terms of a variety of quality of life measures a contributor.39 However, multiple factors can influence at long-term follow-up, SDR did significantly decrease pain or the perception of pain, including comorbidi- the number of subsequent tone-reduction interventions ties, previous treatments received, mental cognition and

J Child Orthop 2018;12:413-427 421 CURRENT CONCEPTS REVIEW: SELECTIVE DORSAL RHIZOTOMY

available social support. In addition, SDR is usually done another concern. The conus approach advocated by Park many years prior to the onset of pain, complicating any and Johnston,16 is preferred by some surgeons to mini- attempts at causal links. Tedroff et al39 measured pain mize this risk compared with multi-level laminectomy, but and pain interference in 18 adults who underwent SDR definitive evidence is lacking. as children. While half reported presence of pain, severity and pain interference were judged to be low compared Energy efficiency (EE) with a general adult CP population, whose SDR status EE during walking is one of the prerequisites of normal was unknown.50 Assessments of pain with CP comparison gait.1 Improving EE of gait can be considered a goal of any groups by Daunter et al46 and Munger et al22 were not able treatment directed at gait improvement. This forms the to show pain improvement attributable to SDR. Therefore, rationale for our routine measurement of EE by oxygen further research is needed to evaluate whether SDR can consumption before and after interventions. Poor EE is a minimize or delay the development of pain in individuals common finding in patients with CP.57 The co-spastic and with CP. prolonged electrical activity associated with spasticity are theoretically additive. An earlier, unmatched study from Joint and musculotendinous contractures our institution suggested that overall, EE improved fol- Although it may improve joint range of movement in lowing SDR.21 However, these results need to be consid- the short-term, evidence exists that SDR does not reduce ered in light of evidence that EE spontaneously decreases the occurrence of joint contractures in the long-term.5,39 with increasing age.58 In our more recent long-term study These findings have been a disappointment. Joint con- with a matched cohort, the improvement in EE for the tractures may be the result of underlying structural soft-­ SDR group approached statistical significance (p = 0.06), tissue abnormalities that are independent of ­muscle tone. while the non-SDR control group did not.22 The findings ­Abnormal lever-arm dysfunction that affects a child’s are far from conclusive, as variations between patients ability to ambulate (e.g. torsional bony abnormalities, were large and sample size was small. Unfortunately, EE foot deformities and joint dysplasia) are also frequently is rarely evaluated and reported in other outcome studies encountered in CP but are neither prevented nor improved on SDR. with SDR. Both joint contractures and lever-arm dysfunc- tion usually present later in childhood and require ortho- paedic surgery. Therefore, the implication is that, for the SDR within a multi-disciplinary, ­ majority of children undergoing SDR, at least one or more multi-modal approach subsequent major surgical intervention is required. A multidisciplinary team working together can help mit- Hip and spine igate the effects of CP on gait and mobility. Clinicians should have a clear understanding of the pros and cons Concerns regarding negative effects of SDR on the hip of alternate management modalities and communicate and spine can affect a clinician’s recommendation regard- together in multi-disciplinary clinics. ing the surgery. At present, there is insufficient evidence to suggest SDR produces any definitive effect. Although SDR and orthopaedic surgery spontaneous improvements in hip dysplasia and sublux- ation have been reported following SDR,52,53 numbers are While SDR is an important tool in tone management, small and no conclusions can be drawn. A recent system- hypertonia is only one aspect of the musculoskeletal atic review into interventions that may prevent hip dys- challenges faced by individuals with CP. Another aspect plasia in CP did not find SDR to have either a positive or is abnormal lever-arm dysfunction, e.g. torsional bony negative effect.54 When hip surgery is required in a child abnormalities, foot deformities and joint dysplasia, which who is otherwise eligible for SDR, timing of such interven- can significantly impede a child’s ability to ambulate. tion is discussed later in this review. These effects are generally seen in later childhood and Reports of postoperative spinal deformity following can only be corrected with orthopaedic surgery. Similarly, SDR have raised concerns. However, many of these con- joint movement limitations due to established contrac- cerns were from earlier studies that included non-ambu- tures do not improve with SDR, nor does SDR prevent the lant children and utilized a more extensive dissection.55,56 development of contractures.5,39 When present, contrac- The high rate of in the general CP population and tures can be improved with orthopaedic intervention. the lack of historical controls compromise ability to inter- In addition to lever arm and joint correc- pret causality in these findings. Experience at our own tion, orthopaedic surgery also plays a role in the tone institution does not support increased incidence of sco- management paradigm (Fig. 1). Transfers of spastic liosis after SDR. Sagittal plane instability following SDR is muscles, e.g. rectus femoris, can improve knee flexion

422 J Child Orthop 2018;12:413-427 CURRENT CONCEPTS REVIEW: SELECTIVE DORSAL RHIZOTOMY

in swing.59 Lengthening of persistently spastic muscles unmask underlying weakness, although children with can be performed, however, at the potential risk of loss significant preoperative weakness in antigravity muscles to muscle strength. Although SDR does not reduce con- should have been excluded from consideration of SDR tractures, it may reduce the need for soft-tissue surgeries (Table 1). The rate at which children recover muscle control that are aimed at decreasing spasticity.22,37 Institutions that correlates with GMFCS level, with those who are GMFCS practice a multi-disciplinary, multi-modal tone manage- I often rehabilitating faster than those that are GMFCS III. ment approach have the opportunity to optimize options Rapid gains can be expected within the first three to six of spasticity control. months following surgery. The goals of physical therapy The relative timing of orthopaedic surgery and SDR also during this time include strengthening (resistance and needs to be considered. If required, orthopaedic surgery endurance), as well as training for a new gait pattern. New is usually indicated at an older age than SDR, when tor- orthoses may be required to promote new walking pat- sional abnormalities and joint contractures are less likely terns. Good motivation, family support, time and cognition to remodel, and also less likely to recur following correc- are crucial throughout the process. These should form part tion. In order to minimize impact on rehabilitation follow- of the preoperative considerations for SDR and should also ing SDR, orthopaedic surgery is not generally performed be communicated to the child and family prior to surgery. until at least one year later, with a few exceptions. If hip subluxation is present and progressive, this may need to SDR and pharmacological tone management be addressed prior to SDR. Currently, it is unclear what Intrathecal baclofen, oral medication, phenol and botu- 54 effect, if any, SDR has on neuromuscular hip dysplasia. linum toxin (BoNT) injections have all been used to alle- However, it is clear that once hip dislocation is established, viate spasticity in CP. A detailed discussion on the pros pain is common and reconstructive options are limited, and cons of each modality is outside the scope of this hence the relative priority this should be given over SDR. review; however, evidence in recent years has renewed Simultaneous SDR and hip reconstruction should be concerns about the efficacy and safety of BoNT injec- avoided, as one operation will restrict the rehabilitation tions. Studies from Korea62 and Australia63 have shown required by the other. that repeated injections resulted in reduced improve- Another situation for earlier orthopaedic management ments in gait and functions. A recent systematic analysis is when an established equinus contracture is present at also suggests that the use of BoNT may cause skeletal time of consideration for SDR. Persistent equinus follow- muscle atrophy.64 ing SDR can significantly affect rehabilitation. If appro- Since SDR eliminates spasticity, the need for anti-spastic priate conservative measures have been ineffective, a injections such as BoNT should be reduced following the low-dose calf-lengthening procedure can be performed procedure. Evidence for this exists.22 In addition, compared concurrently with SDR to allow plantigrade ambulation with those children receiving alternative tone manage- during rehabilitation. Tendoachilles lengthening in zone ment with BoNT and phenol, SDR appears more effective 3 should be avoided, as this can adversely weaken the calf in achieving lasting reduction in tone as measured on the and lead to crouch.60 modified Ashworth scale.22 In light of this evidence and concerns regarding muscle atrophy with repeated BoNT SDR and rehabilitation/physiatry injections, SDR for tone reduction is favoured. Our current Extensive rehabilitation is required and is part of the preference is to use BoNT as a ‘bridging’ treatment for SDR protocol in most institutions performing this proce- very young children, followed by SDR at an appropriate dure.24,38,55,61 Physiatrists and physical therapists should age when gross motor function profile matures and ability be part of the team giving input into the patient selection to participate in rehabilitation improves. process as well as communicating to parents regarding postoperative expectations and rehabilitation plans. At our institution, rehabilitation following SDR differs significantly Cause of suboptimal outcomes following to rehabilitation following SEMLS. Following SDR, patients SDR receive a four- to six-week intensive inpatient programme. Twice-daily physical and occupational therapy commences Suboptimal outcomes following SDR can be divided into on day three post-surgery. After discharge, ongoing out- two broad categories. Those that arise directly as a result of patient therapy continues initially at five days per week, complications from the operation and those that do not. and a tapered programme extends up to one year. Complications During the early postoperative period, a reduction in sensory feedback can limit the ability of the child in main- When rootlet sectioning is not excessive (< 50%) and taining movement control. SDR may also temporarily directed by intraoperative electrical stimulation, results

J Child Orthop 2018;12:413-427 423 CURRENT CONCEPTS REVIEW: SELECTIVE DORSAL RHIZOTOMY

from multiple centres have shown that sensation loss, Table 2 Complications following selective dorsal rhizotomy (SDR). ­Complications seen in our institution following SDR over a ten-year weakness and other neurological compromises are period (Trost et al 200821). All complications were transient and were 22,48,53 rare. Thus, fears of these neurological complica- completed resolved by time of discharge six weeks post-surgery tions following SDR may be unwarranted and can be mitigated. Any neurological problems encountered are Category Number Percentage usually transient. In the earlier years of dorsal rhizotomy Bowel/bladder 11 8 Skin related 9 7 where sectioned rootlets exceeded 50%, and when intra- Wound healing 8 6 operative decisions were not assisted by electrical stimu- Headache 6 4 13,14 Paraesthesia 5 4 lation, greater neurological complications were seen. Weakness 4 3 At our institution, between 25% and 40% of rootlets Miscellaneous related 5 4 are sectioned, as a balance between achieving optimal Miscellaneous unrelated 3 2 spasticity control while minimizing neurological risks. The spasticity reduction seen with this range is similar term, spasticity reduction is evident and modest improve- to those seen in higher percentage rhizotomies, so we ment or maintenance of function can be expected. The see no additional benefit in increasing this percentage. role of SDR may further increase in the future as an alter- A comprehensive review of complications from SDR native to other tone management methods. conducted at our own institution over a ten-year period However, caution should still be exercised as strength revealed that all complications encountered were tran- of evidence in this field remains limited, and some recom- sient and resolved by the time of discharge at six weeks mendations made here rely on institutional experience. post-surgery (Table 2).21 Future improvements in evidence to guide practice can be anticipated through: Other causes 1. collaborative, multi-institutional efforts, such as those 65 Poor outcomes not directly caused by surgical complica- from the Cerebral Palsy Research Network; tions of SDR can be avoided by adhering to the multidisci- 2. the development of improved, quantitative measures plinary team approach. This will minimize the risk of: of spasticity and dystonia; – Poor patient selection. SDR only reduces spasticity, 3. long-term studies with well-matched control groups. and therefore functionally disabling dystonia may be unmasked by the procedure in children with The advent of the above advancements will be welcomed mixed tone. It is also important to identify children and will help us further define the role of SDR in maximiz- with excessive underlying muscle weakness who are ing gait and physical function in children with spastic CP. dependent on high muscle tone for antigravity joint Received 16 July 2018; accepted after revision 04 September 2018. stabilization. – Poor management of lever arm dysfunction. Foot COMPLIANCE WITH ETHICAL STANDARDS deformity, torsional abnormalities of long bones and joint subluxations are very common and adversely FUNDING STATEMENT affect joint moment-generating capacity, leading to No benefits in any form have been received or will be received from a commercial biomechanically inefficient gait. After SDR, close fol- party related directly or indirectly to the subject of this article. low-up with 3DGA is important to monitor for the development of contractures, abnormal torsion and OA LICENCE TEXT crouch gait. Orthoses and orthopaedic surgery are This article is distributed under the terms of the Creative Commons Attribution-Non frequently required to address these. If intervention is Commercial 4.0 International (CC BY-NC 4.0) licence (https://creativecommons.org/ too late, the child may suffer severe deformities and licenses/by-nc/4.0/) which permits non-commercial use, reproduction and distribu- irreversible reduction in function. Routine radiographic tion of the work without further permission provided the original work is attributed. surveillance for hip dysplasia is required to identify sub- luxation that can lead to pain and difficulty walking. ETHICAL STATEMENT Ethical approval: This article does not contain any studies with human participants or animals performed by any of the authors. Summary and future directions Informed consent: Not required for this work. Modern SDR has now been used in our institution as well ICMJE CONFLICT OF INTEREST STATEMENT as others for several decades. With careful patient selec- TFN is on the board of directors for Gillette Children’s Hospital and the Commission on tion guided by 3DGA and a multi-disciplinary approach, Motion Laboratory Accreditation. SDR has been shown to be safe and effective. In the longer All other authors have no declarations to make.

424 J Child Orthop 2018;12:413-427 CURRENT CONCEPTS REVIEW: SELECTIVE DORSAL RHIZOTOMY

ACKNOWLEDGEMENTS 19. Oppenheim WL. Selective posterior rhizotomy for spastic cerebral palsy. A The authors would like to thank Timothy T. Trost for allowing the use of his original review. Clin Orthop Relat Res 1990;253:20-29. image (Figure 2). We would also like to thank Roy Wervey for assistance with 3DGA 20. Grunt S, Fieggen AG, Vermeulen RJ, Becher JG, related graphics. Langerak NG. Selection criteria for selective dorsal rhizotomy in children with spastic cerebral palsy: a systematic review of the literature. Dev Med Child Neurol REFERENCES 2014;56:302-312. 1. Gage J, Schwartz M, Koop S, Novacheck T. The identification and 21. Trost JP, Schwartz MH, Krach LE, Dunn ME, Novacheck TF. treatment of gait problems in cerebral palsy. 2nd ed. London: Mac Keith Press, 2009. Comprehensive short-term outcome assessment of selective dorsal rhizotomy. Dev Med Child 2. Sanger TD, Delgado MR, Gaebler-Spira D, et al. Classification Neurol 2008;50:765-771. and definition of disorders causing hypertonia in childhood. 2003;111:e89-e97. 22. Munger ME, Aldahondo N, Krach LE, Novacheck TF, 3. Nahm NJ, Graham HK, Gormley ME Jr, Georgiadis AG. Schwartz MH. Long-term outcomes after selective dorsal rhizotomy: a retrospective Management of hypertonia in cerebral palsy. Curr Opin Pediatr 2018;30:57-64. matched cohort study. Dev Med Child Neurol 2017;59:1196-1203. 4. Narayanan UG. Management of children with ambulatory cerebral palsy: an 23. Cole GF, Farmer SE, Roberts A, Stewart C, Patrick JH. evidence-based review. J Pediatr Orthop 2012;32:S172-S181. Selective dorsal rhizotomy for children with cerebral palsy: the Oswestry experience. Arch 5. Tedroff K, Löwing K, Jacobson DNO, Åström E. Does loss of Dis Child 2007;92:781-785. spasticity matter? A 10-year follow-up after selective dorsal rhizotomy in cerebral palsy. Dev 24. Graham D, Aquilina K, Cawker S, Paget S, Wimalasundera N. Med Child Neurol 2011;53:724-729. Single-level selective dorsal rhizotomy for spastic cerebral palsy. J Spine Surg 2016;2:195-201. 6. MacWilliams BA, Johnson BA, Shuckra AL, D’Astous JL. 25. Bohannon RW, Smith MB. Interrater reliability of a modified Ashworth Functional decline in children undergoing selective dorsal rhizotomy after age 10. Dev Med scale of muscle spasticity. Phys Ther 1987;67:206-207. Child Neurol 2011;53:717-723. 26. Jethwa A, Mink J, Macarthur C, et al. Development of the Hypertonia 7. Hägglund G, Wagner P. Development of spasticity with age in a total Assessment Tool (HAT): a discriminative tool for hypertonia in children. Dev Med Child Neurol population of children with cerebral palsy. BMC Musculoskelet Disord 2008;9:150. 2010;52:e83-e87. 8. Rosenbaum PL, Walter SD, Hanna SE, et al. Prognosis for 27. Steinbok P, Reiner AM, Beauchamp R, et al. A randomized gross motor function in cerebral palsy: creation of motor development curves. JAMA clinical trial to compare selective posterior rhizotomy plus physiotherapy with 2002;288:1357-1363. physiotherapy alone in children with spastic diplegic cerebral palsy. Dev Med Child 9. Palisano R, Rosenbaum P, Walter S, et al. Development and Neurol 1997;39:178-184. reliability of a system to classify gross motor function in children with cerebral palsy. Dev 28. Wright FV, Sheil EM, Drake JM, Wedge JH, Naumann S. Med Child Neurol 1997;39:214-223. Evaluation of selective dorsal rhizotomy for the reduction of spasticity in cerebral palsy: a 10. D’Aquino D, Moussa AA, Ammar A, Ingale H, Vloeberghs M. randomized controlled trial. Dev Med Child Neurol 1998;40:239-247. Selective dorsal rhizotomy for the treatment of severe spastic cerebral palsy: efficacy and 29. Morris SL, Williams G. A historical review of the evolution of the Tardieu Scale. therapeutic durability in GMFCS grade IV and V children. Acta Neurochir (Wien) 2018;160:811-821. Brain Inj 2018;32:665-669. 11. Steinbok P. Selective dorsal rhizotomy for spastic cerebral palsy: a review. Childs 30. Choi S, Shin YB, Kim SY, Kim J. A novel sensor-based assessment of lower Nerv Syst 2007;23:981-990. limb spasticity in children with cerebral palsy. J Neuroeng Rehabil 2018;15:45. 12. Foerster O. Über eine neue operative Methode der Behandlung spastischer 31. Himmelmann K, Hagberg G, Wiklund LM, Eek MN, Lähmungen mittels Resektion hinterer Rückenmarkswurzeln. Z Orthop Chir 1908;22:463-474. Uvebrant P. Dyskinetic cerebral palsy: a population-based study of children born 13. Gros C, Ouaknine G, Vlahovitch B, Frèrebeau P. Selective between 1991 and 1998. Dev Med Child Neurol 2007;49:246-251. posterior radicotomy in the neurosurgical treatment of pyramidal hypertension. 32. Marsico P, Frontzek-Weps V, Balzer J, van Hedel HJA. Neurochirurgie 1967;13:505-518. Hypertonia assessment tool: reliability and validity in children with neuromotor disorders. 14. Fasano VA, Broggi G, Barolat-Romana G, Sguazzi A. Surgical J Child Neurol 2017;32:132-138. treatment of spasticity in cerebral palsy. Childs Brain 1978;4:289-305. 33. Beattie C, Gormley M, Wervey R, Wendorf H. An electromyographic 15. Peacock WJ, Arens LJ. Selective posterior rhizotomy for the relief of spasticity protocol that distinguishes spasticity from dystonia. J Pediatr Rehabil Med 2016;9:125-132. in cerebral palsy. S Afr Med J 1982;62:119-124. 34. Roberts A, Stewart C, Freeman R. Gait analysis to guide a selective 16. Park TS, Johnston JM. Surgical techniques of selective dorsal rhizotomy for dorsal rhizotomy program. Gait Posture 2015;42:16-22. spastic cerebral palsy. Technical note. Neurosurg Focus 2006;21:e7. 35. Trost JP. Physical assessment and observational gait analysis. In: Gage JR, ed. The 17. Park TS. Selective dorsal rhizotomy: an excellent therapeutic option for spastic treatment of gait problems in cerebral palsy. London: Mac Keith Press, 2004:71-89. cerebral palsy. Clin Neurosurg 2000;47:422-439. 36. Schwartz MH, Rozumalski A, Steele KM. Dynamic motor control is 18. Abbott R, Forem SL, Johann M. Selective posterior rhizotomy for the associated with treatment outcomes for children with cerebral palsy. Dev Med Child Neurol treatment of spasticity: a review. Childs Nerv Syst 1989;5:337-346. 2016;58:1139-1145.

J Child Orthop 2018;12:413-427 425 CURRENT CONCEPTS REVIEW: SELECTIVE DORSAL RHIZOTOMY

37. Dudley RWR, Parolin M, Gagnon B, et al. Long-term functional benefits 52. Hicdonmez T, Steinbok P, Beauchamp R, Sawatzky B. of selective dorsal rhizotomy for spastic cerebral palsy. J Neurosurg Pediatr 2013;12:142-150. Hip joint subluxation after selective dorsal rhizotomy for spastic cerebral palsy. J Neurosurg 38. Josenby AL, Wagner P, Jarnlo GB, Westbom L, Nordmark E. 2005;103:10-16. Motor function after selective dorsal rhizotomy: a 10-year practice-based follow-up study. 53. Nordmark E, Josenby AL, Lagergren J, et al. Long-term outcomes Dev Med Child Neurol 2012;54:429-435. five years after selective dorsal rhizotomy.BMC Pediatr 2008;8:54. 39. Tedroff K, Löwing K, Åström E. A prospective cohort study investigating 54. Miller SD, Juricic M, Hesketh K, et al. Prevention of hip gross motor function, pain, and health-related quality of life 17 years after selective dorsal displacement in children with cerebral palsy: a systematic review. Dev Med Child Neurol rhizotomy in cerebral palsy. Dev Med Child Neurol 2015;57:484-490. 2017;59:1130-1138. 40. Bax M, Tydeman C, Flodmark O. Clinical and MRI correlates of cerebral 55. Golan JD, Hall JA, O’Gorman G, et al. Spinal deformities following palsy: the European Cerebral Palsy Study. JAMA 2006;296:1602-1608. selective dorsal rhizotomy. J Neurosurg 2007;106:441-449. 41. Volpe JJ. Brain injury in premature infants: a complex amalgam of destructive and 56. Johnson MB, Goldstein L, Thomas SS, et al. Spinal deformity developmental disturbances. Lancet Neurol 2009;8:110-124. after selective dorsal rhizotomy in ambulatory patients with cerebral palsy. J Pediatr Orthop 42. Peacock WJ, Arens LJ, Berman B. Cerebral palsy spasticity. Selective 2004;24:529-536. posterior rhizotomy. Pediatr Neurosci 1987;13:61-66. 57. Rose J, Gamble JG, Burgos A, Medeiros J, Haskell WL. 43. McLaughlin J, Bjornson K, Temkin N, et al. Selective dorsal rhizotomy: Energy expenditure index of walking for normal children and for children with cerebral palsy. meta-analysis of three randomized controlled trials. Dev Med Child Neurol 2002;44:17-25. Dev Med Child Neurol 1990;32:333-340. 44. McLaughlin JF, Bjornson KF, Astley SJ, et al. Selective dorsal 58. Bolster EAM, Balemans ACJ, Brehm M-A, Buizer AI, rhizotomy: efficacy and safety in an investigator-masked randomized clinical trial. Dev Med Dallmeijer AJ. Energy cost during walking in association with age and body height Child Neurol 1998;40:220-232. in children and young adults with cerebral palsy. Gait Posture 2017;54:119-126. 45. Grunt S, Becher JG, Vermeulen RJ. Long-term outcome and adverse 59. Lee SY, Kwon SS, Chung CY, et al. Rectus femoris transfer in cerebral effects of selective dorsal rhizotomy in children with cerebral palsy: a systematic review.Dev palsy patients with stiff knee gait.Gait Posture 2014;40:76-81. Med Child Neurol 2011;53:490-498. 60. Borton DC, Walker K, Pirpiris M, Nattrass GR, Graham HK. 46. Daunter AK, Kratz AL, Hurvitz EA. Long-term impact of childhood Isolated calf lengthening in cerebral palsy. Outcome analysis of risk factors. J Bone Joint Surg selective dorsal rhizotomy on pain, fatigue, and function: a case-control study. Dev Med Child [Br] 2001;83-B:364-370. Neurol 2017;59:1089-1095. 61. Langerak NG, Tam N, Vaughan CL, Fieggen AG, Schwartz 47. Bolster EAM, van Schie PEM, Becher JG, et al. Long-term MH. Gait status 17-26 years after selective dorsal rhizotomy. Gait Posture 2012;35:244-249. effect of selective dorsal rhizotomy on gross motor function in ambulant children with 62. Hong BY, Chang HJ, Lee SJ, et al. Efficacy of repeated botulinum toxin spastic bilateral cerebral palsy, compared with reference centiles. Dev Med Child Neurol type a injections for spastic equinus in children with cerebral palsy—a secondary analysis of 2013;55:610-616. the randomized clinical trial. Toxins (Basel) 2017;9:253. 48. Langerak NG, Lamberts RP, Fieggen AG, et al. A prospective 63. Read FA, Boyd RN, Barber LA. Longitudinal assessment of gait quality in gait analysis study in patients with diplegic cerebral palsy 20 years after selective dorsal children with bilateral cerebral palsy following repeated lower limb intramuscular Botulinum rhizotomy. J Neurosurg Pediatr 2008;1:180-186. toxin-A injections. Res Dev Disabil 2017;68:35-41. 49. Subramanian N, Vaughan CL, Peter JC, Arens LJ. Gait 64. Mathevon L, Michel F, Decavel P, et al. Muscle structure and stiffness before and 10 years after rhizotomy in children with cerebral palsy spasticity. J Neurosurg assessment after botulinum toxin type A injection. A systematic review. Ann Phys Rehabil 1998;88:1014-1019. Med 2015;58:343-350. 50. Schwartz L, Engel JM, Jensen MP. Pain in persons with cerebral palsy. 65. Stevenson RD. Integration of research and clinical practice: the future is now. Dev Arch Phys Med Rehabil 1999;80:1243-1246. Med Child Neurol 2017;59:119-120. 51. Hilberink SR, Roebroeck ME, Nieuwstraten W, et al. Health 66. Kim HS, Steinbok P, Wickenheiser D. Predictors of poor outcome issues in young adults with cerebral palsy: towards a life-span perspective. J Rehabil Med after selective dorsal rhizotomy in treatment of spastic cerebral palsy. Childs Nerv Syst 2007;39:605-611. 2006;22:60-66.

426 J Child Orthop 2018;12:413-427 CURRENT CONCEPTS REVIEW: SELECTIVE DORSAL RHIZOTOMY

APPENDIX A

Selection criteria for selective dorsal rhizotomy (SDR) at other centres

Oswestry* Great Ormond Street† History Age: 5 to 10 years Age: 3 to 14 years Absence of chronic conditions e.g. epilepsy Minimum 6 months since last BoTN injection Intelligence Quotient (IQ) > 70 Minimum 6 months since orthopaedic surgery Well-motivated, emotionally robust Cognitive and emotionally suitable No previous multi-level surgery Supportive home environment Supportive home environment Access to rehabilitation facilities

Examination or severe hemiplegia with no significant Spastic diplegia with no significant ataxia or dystonia or dystonia Moderate to severe spasticity Typically Gross Motor Function Classification System II or III Mean lower limb power > 3 on Medical Research Council scale Good trunk control and lower extremity antigravity strength At least moderate movement control No significant scoliosis At least moderate balance Absence of severe fixed joint deformity No significant scoliosis

Gait analysis Not included Not included

Imaging No injury to basal ganglia No injury to basal ganglia, brainstem or cerebellum No hip dysplasia Riemer’s index < 40% Not excessively high body mass index

*from Oswestry: Robert Jones and Agnes Hunt Orthopedic Hospital1 †from Great Ormond Street Hospital2

REFERENCES 2. Graham D, Aquilina K, Cawker S, Paget S, Wimalasundera N. 1. Cole GF, Farmer SE, Roberts A, Stewart C, Patrick JH. Single-level selective dorsal rhizotomy for spastic cerebral palsy. J Spine Surg 2016;2:195-201. Selective dorsal rhizotomy for children with cerebral palsy: the Oswestry experience. Arch Dis Child 2007;92:781-785.

J Child Orthop 2018;12:413-427 427